Antibacterial Activity, Cytotoxicity and Chemical Constituents of Hydnora Johannis Roots

Available online at www.sciencedirect.com

South African Journal of Botany 78 (2012) 228–234 www.elsevier.com/locate/sajb

Antibacterial activity, cytotoxicity and chemical constituents of Hydnora johannis roots

S. Yagi a,b, F. Chrétien a, R.E. Duval c, S. Fontanay c, M. Maldini d, S. Piacente d, M. Henry a, ⁎ Y. Chapleur a, D. Laurain-Mattar a,

a Groupe S.U.C.R.E.S., UMR 7565 CNRS-Nancy-Université, BP 239,54506 Nancy-Vandoeuvre, France b Botany Department, Faculty of Science, University of Khartoum, P.O. Box 321, Khartoum, Sudan c GEVSM, UMR 7565 CNRS, Nancy-Université, Nancy, France d Università degli Studi di Salerno, Dipartimento di Scienze Farmaceutiche, via Ponte don Melillo, 84084 Fisciano (Sa), Italy

Received 17 March 2011; received in revised form 8 September 2011; accepted 14 September 2011

Abstract

In Sudan, the roots of Hydnora johannis (Hydnoraceae) are traditionally used for the treatment of dysentery, diarrhea, cholera and swelling tonsillitis. The ethnomedicinal value of H. johannis was investigated through phytochemical study, in vitro antibacterial activity and preliminary cytotoxic tests. Determination of total phenols, flavonoids and proanthocyanidins was carried out using spectrometric methods. The antibacterial activity of the water and ethanolic extracts was determined using the microdilution method. Pure compounds were isolated from the ethyl acetate extract by chromatographic methods and their structures were established by spectroscopic methods. Cytotoxicity assay was performed against selected human mouth epidermoid carcinoma cell line (KB), and non-cancer human fetal lung cell line (MRC-5). Both water and ethanol (70%) extracts were found to contain the same amount of total phenols and proanthocyanidins, whereas the level of flavonoids was higher in the ethanol extract. The water extract was found to possess antibacterial activity against Enterococcus faecalis (MIC value of 16 μg/mL), Bacillus subtilis, B. cereus and Staphylococcus aureus (MIC values of 64 μg/mL) but not against bacteria mainly responsible for diarrhea. This leads to the suggestion that, the mode of action of water extract which is rich in tannins was not connected to their inhibition to the diarrhea bacteria but to their action on the digestive tract. Reduction in potency of the water and ethanol (70%) extracts when fractionated was observed. The ethyl acetate fraction obtained from fractionation of ethanol extract possessed only activity against the two strains of S. aureus with MIC values of 128 μg/mL. In addition, six compounds were isolated from the ethyl acetate extract as cirsiliol (3′,4′,5-trihydroxy-6-7-dimethoxy flavone) (1), trans 3′5-dihydroxy-4′7-dimethoxydihydroflavonol (2), oleic acid (3), vanillin (4-hydroxy-3-methoxybenzaldehyde) (4), protocatechuic acid (3,4 dihydroxy benzoic acid) (5) and dl catechin (trans (+) 2-(3,4-dihydroxyphenyl-3,4-dihydro-2H-1-benzopyran-3,5,7-triol)) (6). Four compounds from the ethyl acetate extract were also identified by GCMS as stigmasterol (7), oleic acid (3), myristic acid (8), and palmitic acid (9). Little cytotoxicity is reported against the cell lines used. Thus, the safety of this plant in the traditional medicine should be verified by much further testing, including in vivo experiments and clinical studies. © 2011 SAAB. Published by Elsevier B.V. All rights reserved.

Keywords: Antibacterial activity; Chemical constituents; Cytotoxicity; Hydnora johannis

1. Introduction reported that diarrhea is the highest course of infant mortal- ity disease in the world, occurring mostly in developing Diarrhea has long been recognized as one of the most im- countries (Murray et al., 2001). It is a major health concern portant health problems in developing countries. It has been in developing countries and remains an important clinical problem even in developed countries despite improvements ⁎ in public health and economic wealth (Casburn-Jones and Corresponding author. – E-mail address: [email protected] Farthing, 2004). It is estimated that during the next 20 (D. Laurain-Mattar). 30 years, diarrhea along with other infectious diseases will

0254-6299/$ - see front matter © 2011 SAAB. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.sajb.2011.09.010 S. Yagi et al. / South African Journal of Botany 78 (2012) 228–234 229 remain a cause of global health concern (Meyrowitsch and 2. Material and methods Bygbjerg, 2007). Treatment of diarrhea is generally non- specific and is usually aimed at reducing the discomfort and 2.1. Plant material inconvenience of frequent bowel movements (Brunton, 1996). To overcome the menace of diarrheal diseases in devel- Roots of H. johannis were collected from Aldamasin district of opingcountries,theWorldHealthOrganization(WHO)has The Blue Nile Region, Sudan in January 2006. The plant was iden- included a program for the control of diarrhea, which involves tified and a voucher specimen no 41936HBD was deposited in the the use of traditional herbal medicine (Atta and Mouneir, Herbarium of Botany Department, Faculty of Science, University 2004; Snyder and Merson, 1982). Therefore, it is necessary of Khartoum for future reference. to establish the scientific basis for the therapeutic actions of traditional plant medicines. Several plants have been reported 2.2. Preparation of extracts to be used in treating and managing diarrheal diseases (Agunu et al., 2005). Traditional medical practitioners play important Two main types of extracts (water and ethanol) were prepared roles in health care delivery in Sudan. With recent increasing and then were further subjected to different fractionations. Water interest in alternative/herbal medicine for the prevention and extract (WE) was prepared by simple maceration of 500 g of treatment of various illnesses, there is an increasing concern powdered roots in 1500 mL of distilled water maintained at am- about the safety of medicinal plants. In Sudan, toxicoses bient temperature for 4 h. Extract was first filtered on filter with medicinal plants are encountered among the village paper, then centrifuged for 15 min at 0.704×g and finally populations who are dependent mainly on traditional medi- freeze-dried to yield 13.56 g of WE. In order to get water fraction cine. Most cases of poisoning result from over dosages be- not containing tannins, the method of tannin separation using rab- cause, in general, there is no standardized dosage system in bit leather described in the European Pharmacopeia (edition traditional medical practice (Saad et al., 2006). To ensure re- 2005) was adopted. Two grams of WE was dissolved in producible quality of herbal products, preliminary validation 200 mL distilled water. Two grams of rabbit leather powder (pur- in laboratory and an optimization of their formulation are ut- chased as substance R according to the purity criteria mentioned most essential. in the European Pharmacopeia) was added and the mixture was The Hydnoraceae with only two genera Hydnora and put under agitation for 1 h at room temperature, and then filtered. Prosopanche include some of the strangest plants in the To insure the absence of tannins from the filtrate, 200 μLof world. The vegetative plant body is highly reduced consist- FeCl3 (10%) were added to 1 mL of filtrate. A dark violet color ing of only roots, flowers and fruits. In fact, the Hydnora- is developing in the presence of tannins. The filtrate was repeat- ceae are the only known angiosperm with no leaves or edly extracted by rabbit leather powder until tannins were absent. scales of any sort (Musselman, 1991). There are currently Filtrate free of tannins was freeze dried to yield 0.24 g of non- five species recognized; H. johannis Becc. (=H. abyssinica tannins water extract (NTWE). Tannins fraction (TF) was sepa- A. Br. =H. solmsiana Dinter), H. africana Thunb., H. escu- rated from rabbit leather powder by adding 5 g of trypsin to the lenta Jum. and H. Perrier, H. triceps Drège and E. Mey. and complex (tannins+rabbit leather) under light agitation at 37 °C H. sinandevu Beentje and Q. Luke (Tennakoon et al., 2007). for 36 h. The supernatant was extracted with phenol:chloro- H. byssinica occurs from northern Namibia, across most of form:ethanol (25:24:1 (v/v/v)) to get rid of the protein and then Angola and the DRC, to Ethiopia, Sudan and the Arabian freeze dried to yield 9.94 g of TF. The ethanolic extract was pre- peninsula (Bolin et al., 2005; Musselman and Visser, pared by soaking 2 kg of roots powder in 70% ethanol at ambient 1989). Moreover, recent studies demonstrated its occurrence temperature for 5 days. The extract was decanted, filtered under in South Africa (Williams et al., 2011a), as well as Mozam- vacuum, concentrated in a rotary evaporator, then extracted suc- bique and Malawi (Williams et al., 2011a, 2011b). H. johan- cessively with n-hexane (250 mL×3), CHCl3 (300 mL×3), and nis is found parasitizing the roots of Acacia nilotica EtOAc (300 mL×3) and were evaporated under reduced pressure (Musselman and Visser, 1987). The plants are reportedly to yield 1.16 g, 3.50 g and 5.60 g residues respectively. The water used to treat diarrhea, piles, acne, menstrual problems, stom- extract (70.6 g) left was lyophilized. ach cramps and to stop bleeding (Dold and Cocks, 2003; Hutchings et al., 1996; Musselman and Visser, 1989). In 2.3. Determination of total phenolics Sudan, decoction of the roots of H. johannis is traditionally used for the treatment of dysentery, diarrhea, cholera and Water and ethanol extracts were resuspended separately in swelling tonsillitis (El Ghazali, 1997). Phytochemical stud- ethanol to make 50 mg/mL stock solutions. Total phenol contents ies and biological evaluation of this plant are scanty. Con- in the extracts were determined using modified Folin–Ciocalteu sidering the widespread use of H. johannis in some parts method (Wolfe et al., 2003). An aliquot of the extract was of Sudan, the present study was conducted to assess the pos- mixed with 5 mL Folin–Ciocalteu reagent (previously diluted sible antibacterial activity and toxic effect of the roots of with water at 1:10 v/v) and 4 mL (75 g/L) of sodium carbonate. H. johannis water extract (the conventional form in which The tubes were vortexed for 15 s and allowed to stand for the plant is administered in traditional medicine) and ethano- 30 min at 40 °C for color development. Absorbance was then lic extract and to get information of the amount of phenolics measured at 765 nm using the SHIMADZU UV-2550 UV-VS and major constituents of the roots. spectrophotometer. Sample extracts were evaluated at a final 230 S. Yagi et al. / South African Journal of Botany 78 (2012) 228–234 concentration of 0.1 mg/mL. Total phenolic contents were After 24 h of growth, suspensions were diluted in distilled expressed as tannic acid (purchased from ROTH) equivalents in water to obtain a final inoculum of 5×105–5×106 CFU/mL. milligram per gram sample (mg/g). Purity of isolates was checked throughout the study by exami- nation of colony morphology and Gram stain procedure. The 2.4. Determination of total flavonoids test samples were first of all dissolved in distilled water for water and ethanol extracts and in DMSO for other test samples. Estimation of the total flavonoids in the plant extracts was Two-fold serial dilutions of each extract/fraction/pure com- carried out using the method of Ordon Ez et al. (2006).To pound were prepared in Mueller Hinton Broth in 96-wells 0.5 mL of sample, a volume of 0.5 mL of 2% AlCl3 ethanol so- plate (Greiner, ref 65061, Strasbourg, France), starting from a lution was added. After 1 h at room temperature, the absor- stock solution of 1024 mg/L. An equal volume of bacterial in- bance was measured at 420 nm. A yellow color indicated the oculum was added to each well on the microtiter plate contain- presence of flavonoids. Extract samples were evaluated at a ing 0.05 mL of the serial extract/fraction/compound dilutions. final concentration of 0.1 mg/mL. Total flavonoid content was After incubation for 18–24 h at 35 °C, the MICs were deter- calculated as quercetin (purchased from ROTH) equivalents mined with a ELISA reader (read at 540 nm, Multiskan EX, in milligram per gram sample (mg/g). Thermo Electron Corporation, France) as the lowest concentration of compound which absorbance was comparable with the 2.5. Determination of total proanthocyanidins negative control wells (broth only or broth with test sample, without inoculum). Also, concentrations of extracts inhibiting Determination of proanthocyanidins was based on the proce- 50% of the bacterial growth (MIC50) were graphically deter- dure reported by Sun et al. (1998). A volume of 0.5 ml of mined, using the positive control (i.e. broth with inoculum, 0.1 mg/mL of extract solution was mixed with 3 mL of 4% van- but without test sample) as the 100% of growth. Results were illin–methanol solution and 1.5 mL hydrochloric acid; the mix- expressed as means of four independent determinations. Refer- ture was allowed to stand for 15 min. The absorbance was ence agents in clinical use were also tested in order to check the measured at 500 nm. Extract samples were evaluated at a susceptibility of strains: amoxicillin (Sigma, A8523-1G), peni- final concentration of 0.1 mg/mL. Total proanthocyanidin con- cillin G (Sigma, 13752-1G), ticarcillin (Sigma, T5639-1G) and tent was expressed as catechin (purchased from ROTH) equiv- vancomycin (Sigma, V2002-1G). alents in milligram per gram sample (mg/g). 2.7. Cytotoxicity assay 2.6. Antibacterial activity The human cell lines KB (derived from mouth epidermoid car- 2.6.1. Bacteria cinoma) and MRC5 (derived from non-cancer human fetal lung) Escherichia coli ATCC 25922, S. aureus ATCC 25923 and were obtained from ECACC (Salisbury, UK) and grown in Dul- ATCC 29213, Enterococcus faecalis ATCC 29212, Pseudomonas becco's modified Eagle's medium supplemented with 25 mM aeruginosa ATCC 27853, B. subtilis ATCC 6633 and Klebsiella glucose, 10% (v/v) fetal calf serum, 100 UI penicillin, oxytoca ATCC 700324 were used as reference strains, following 100 μg/mL streptomycin and 1.5 μg/mL fungizone and kept guidelines of the Clinical and Laboratory Standards Institute under 5% CO2 at 37 °C. Cells were plated in 96-well tissue cul- [CLSI, formerly National Committee for Clinical Laboratory Stan- ture microplates at a density of 650 cells/well in 200 μLmedium dards (NCCLS, 2003)] and of the Comité de l'Antibiogramme de and treated 24 h later with extract/compound dissolved in DMSO la Société Française de Microbiologie (CA-SFM, 2008). All other with compound concentrations 1 or 10 μM (or 1 and 10 μg/mL strains (Corynebacterium jeikeium S1, Listeria monocytogenes S1, for extracts) using a Biomek 3000 automate (Beckman–Coulter). E. coli O55 S1, Proteus vulgaris S1, Salmonella enterica Anatum Each concentration was tested on three replicate wells. Taxotere® S1, S. enterica Enteritidis S1, S. enterica Enteritidis S2, S. enterica (control compound) received the same volume of DMSO. After Newport, S. enterica Typhimurium S1, S. enterica Typhimurium 72 h exposure MTS reagent (Promega) was added and incubated S2, S. enterica Typhimurium S3, Shigella boydii S1, S. flexneri for 3 h at 37 °C, the absorbance was monitored at 490 nm and re- S1, S. sonnei S1, Yersinia enterocolitica S1, Y. enterocolitica S2) sults expressed as the inhibition of cell proliferation calculated as were the clinical isolates of bacterial species of interest (i.e. that the ratio [(1-(OD490 treated/OD490 control)) 100]. IC50 calculat- can be associated with dysenteric or diarrheic syndromes) selected ed as the concentration of extract eliciting a 50% inhibition of cell from the ABC Platform Bugs Bank, Nancy, France. Strains were proliferation on using different concentrations ranging from 50 to grown on Mueller Hinton Agar (Difco, 225250) or Mueller Hinton 0.003 μg/mL in a fixed volume of DMSO (1% final concentra- Broth (Difco, 275730). tion). The tests were performed in duplicate.

2.6.2. Minimum inhibitory concentrations (MIC) determination 2.8. Isolation and characterization of pure compounds MICs were determined by broth microdilution method recommended by the CLSI (NCCLS, 2003) as previously de- 2.8.1. General experimental procedures scribed (Fontanay et al., 2008). For MIC determination; sus- Column chromatography (CC) was carried out on a silica gel pensions were prepared by suspending one isolated colony (70–230 mesh) column or on flash chromatography using glass from Mueller Hinton plates in 5 mL of Mueller Hinton Broth. Kontes Chromaflex columns packed with silica gel of particle S. Yagi et al. / South African Journal of Botany 78 (2012) 228–234 231 size (40–63 μm) or with pre packed cartridges. Gradient system Table 1 is generated by 2 pumps Gilson model 306. Thin layer chroma- Phenol contents of the water and ethanol extracts from Hydnora johannis roots. tography (TLC) was performed on pre-coated silica gel plates Phenolics Water extract Ethanol extract (70%) a (DC-Alufolien 60F254 of E. Merck), and spots were located by Total polyphenols 662.10±14.4 661.58±43.9 b using vanillin/H2SO4 spraying reagent. GCMS analysis was car- Flavonoids 25.75±0.72 31.82±0.31 c ried out on QP 20–10 Shimadzu GC-MS equipment. Supelco eq- Proanthocyanidins 44.54±1.40 44.72±1.52 uity 1 column with a film thickness of 30 m×0.25 μm was used. (n=3, X±SEM). A Wiley 229 library search was conducted on major peaks of the a Expressed as mg tannic acid/g of dry plant material. b sample in order to identify the components of the sample. IR Expressed as mg quercetin/g of dry plant material. c Expressed as mg catechin/g of dry plant material. spectra were recorded on JASCO 302-A spectrophotometer. EI- MS spectra were recorded on Finnegan MAT 311 mass spectrometer with a MASPECO Data System and HR-MS spectra approximately the same amount of phenols and proanthocyani- were recorded in ESI with a Bruker microTOFQ. All NMR exper- dins. In fact, little is known about the chemistry of the genus Hyd- iments were performed with a Bruker Avance DRX-400 instru- nora. A preliminary test of the presence of secondary metabolites ment operating at a proton frequency of 400.13 MHz. This in H. johannis roots indicated the presence of tannins and phenols spectrometer was equipped with a 5 mm broadband inverse de- in high level concentration. Other constituents were found in tection z-gradient probe tuned to 13C (100.61 MHz). For all 1D moderate concentration (Saadabi and Ayoub, 2009). Interesting- and 2D NMR experiments pulse sequences provided by the spec- ly, the host plant A. nilotica (L.) Wild ex Del. parasitized by H. trometer manufacturer were used. johannis was reported by many investigators to be also very rich in tannins and phenols (Ayoub, 1984; Cannas, 1999; 2.8.2. Isolation of pure compounds Mahdi et al., 2006). Pods, stem bark and leaves of A. nilotica The EtOAc fraction (3.3 g) was subjected to column chroma- are also used for the treatment of diarrhea in the traditional med- tography eluted using a gradient of petroleum ether (P.E), icine of many African and Asian countries (Agunu et al., 2005; 500 mL followed by EtOAc:MeOH (19:1, 500 mL; 11.5:1, Ambasta, 1994). 1000 mL; 7:1, 1500 mL and 1:1, 500 mL). A total number of 80 fractions (50 mL each) were collected and finally 8 fractions 3.2. Antibacterial activity were obtained on combining the eluates according to their similarity behavior on TLC. Fraction 1 (11 mg) eluted with P.E: Results of MIC and MIC50 for the water extract of H. johannis EtOAc (9:1) was submitted to GCMS analysis. Fractions 2 roots against several human bacterial pathogens were presented (551 mg), 4 (50 mg), and 5 (122 mg) were applied repeatedly to in Table 2. The water extract showed a significant activity against silica gel CC and eluted with P.E:EtOAc or CHCl3:MeOH gradi- E. faecalis with MIC values of 16 μg/mL and on both strains of S. ents to afford pure compounds (1) (2 mg) in CHCl3:MeOH (2%); aureus, B. subtilis and B. cereus with MIC values of 64 μg/mL. (2) (9 mg) in P.E:EtOAc (8:2), (3) (15 mg) in P.E:EtOAc (9:1) The water extract of H. johannis was devoid of any antibacterial and (8:2) and (4) (13 mg) in P.E:EtOAc (9:1). For the next col- activity against all other bacterial strains tested in particular, those umn 2.1 g of the EtOAc extract was subjected to flash chroma- are mainly responsible of diarrhea i.e. Escherichia spp., Salmo- tography using a gradient of hexane to EtOAc (4000 mL) as nella spp., Shigella spp., or Yersinia spp. Interestingly, the eluent. A total number of 200 fractions (20 mL each) were col- water fraction showed MIC50 at concentrations of 128 and lected and finally 8 fractions were obtained on combining the el- 156 μg/mL against E. coli and two Salmonella spp respectively. uates according to their similarity behavior on TLC. Fractions 4 In fact, all Gram-negative bacteria tested were resistant to the (445 mg) and 5 (314 mg) were subjected to repeated flash chro- test extract. The reasons for the differential sensitivity pattern be- matography using hexane:acetone mixture of increasing polarity tween Gram-positive and Gram-negative bacterial strains could as eluent to obtain compounds (5) (75 mg) and (6) (538 mg). be due to the outer phospholipids membrane with the structural lipopolysaccharide components, which make their cell wall im- 3. Results and discussion penetrable to antimicrobial agents (Nikaido, 1994), while the Gram-positive bacteria is more susceptible having only an outer 3.1. Determination of total phenols peptidoglycan, which is not an effective permeability barrier (Scherer and Gerhardt, 1971). Several studies have shown the ac- The total phenolic contents of the water and ethanol (70%) ex- tion of tannins for the treatment of diarrhea and cholera (Devi et tracts of H. johannis roots were found to be 662.10 and al., 2002; Miranda et al., 1993; Yu et al., 2000). Tannins and fla- 661.58 mg tannic acid equivalent/g of dry sample. Results also vonoids from the leaves of Eremomastax speciosa have been revealed that the ethanol extract has higher levels of flavonoid shown to possess antidiarrheal activity (Oben et al., 2006). The contents than the water extract with values of 31.82 and high tannin concentration in the Hydnora spp. rhizomes imparts 25.75 mg quercetin equivalent/g of dry sample respectively. a strong astringency and bitterness, and this may be the reason The proanthocyanidins were found to be 44.54 and 44.72 mg cat- for its efficacy in treating intestinal ailments (Dold and Cocks, echin equivalent/g of dry sample in the same extracts (Table 1). It 2003; Musselman, 1984; Musselman and Visser, 1989). Howev- was clear that both extracts contain significant amount of phe- er, in this study it was suggested that, the curing potency of the nols, flavonoids and proanthocyanidins and they contain roots H. johannis was not mainly associated with the presence 232 S. Yagi et al. / South African Journal of Botany 78 (2012) 228–234

Table 2 Table 3 Antibacterial activity of water extract from Hydnora johannis roots expressed Antibacterial activity of extracts from Hydnora johannis roots expressed as as MICs and MIC50 (μg/mL). MICs (μg/mL).

Bacterial strains Gram MIC MIC50 Extracts E. faecalis E. coli S. aureus S. aureus 2 P. aeruginosa (μg/mL) (μg/mL) WE 16 N256 64 64 N256 Bacillus cereus S1 + 64 − NTWE N256 N256 N256 N256 N256 Bacillus subtilis ATCC 6633 + 64 − TF N256 N256 N256 N256 N256 Corynebacterium jeikeium S1 + N256 N256 WE, water; NTWE, water not containing tannin; TF, tannins fraction, MICs Enterococcus faecalis ATCC 29212 + 16 values are representative of three independent determinations. Listeria monocytogenes S1 + N256 N256 Staphylococccus aureus ATCC 25923 + 64 − Staphylococccus aureus ATCC 29213 + 64 − − N − Escherichia coli ATCC 25922 256 synergistically to produce a particular therapeutic effect, their Escherichia coli O55 S1 − N256 128 Klebsiella oxytoca ATCC 700324 − N256 N256 separation may lead to a loss of the desired activity. In fact, syn- Proteus vulgaris S1 − 256 − ergism can be due to the individual action of different constitu- Pseudomonas aeruginosa − N256 − ents present in the extract at multiple target sites/parameters or ATCC 27853 by having an additive effect against a single target site/parameter − N Salmonella enterica Anatum S1 256 256 (Ryabushkina, 2005). These results suggested that the tannins Salmonella enterica Enteritidis S1 − N256 N256 Salmonella enterica Enteritidis S2 − N256 256 could be implicated in the antibacterial activity of H. johannis Salmonella enterica Newport − N256 N256 against E. faecalis and S. aureus but, their activity was associated Salmonella enterica Typhimurium S1 − N256 N256 with the presence of other compounds found in the water extract. Salmonella enterica Typhimurium S2 − N256 N256 The sensitivity of E. faecalis, E. coli, S. aureus and P. aeruginosa − N N Salmonella enterica Typhimurium S3 256 256 to the ethanol (70%) extracts and its fractions (i.e. hexane, chloro- Shigella boydii S1 − N256 N256 Shigella flexneri S1 − N256 N256 form ethylacetate and water fractions) as well as pure compounds Shigella sonnei S1 − N256 N256 (catechin, vanillin and protocatechuic acid) obtained from the Yersinia enterocolitica S1 − N256 N256 ethyl acetate fraction was carried out (Table 4). The ethanol Yersinia enterocolitica S2 − N256 N256 (70%) extract showed significant activity against E. faecalis Amoxicillin* 4 μ b (MIC value of 16 g/mL) and S. aureus strains (MIC value of Penicillin G* 1 μ Ticarcillin* 16 64 g/mL). Reduction in potency of the ethanol (70%) extracts Vancomycin* b1 when fractionated was observed. Only the ethyl acetate and *Reference antibiotics; MICs and MIC50 values are representative of four inde- water fractions possessed activity against the two strains of pendent determinations. S. aureus with MIC values of 128 μg/mL. Flavonoids from Marchantia convolute leaves (Xiao et al., 2005) and tannins like tannic acid (Akiyama et al., 2001)havebeenreportedtoin- of antibacterial activity agent(s) against bacterial species respon- hibit the growth of S. aureus and could be responsible for the ob- sible of dysentery or diarrhea. This might be attributed to the role served antistaphylococcal activity. However, in this study the of tannins in reducing the effect through denaturing the proteins pure compounds isolated from the ethyl acetate fraction did not by the formation of protein tannate, thereby causing the intestinal show any activity against S. aureus. mucosa to become more resistant and by stimulating the normal- A previous study on antibacterial activity of the water, metha- ization of the deranged water transport across the mucosal cells nol and chloroform extracts of H. johannis roots against S. aureus, and the reduction of the intestinal transit, which have more useful B. subtilis, E. coli and P. aeruginosa, using the cup-plate agar therapeutic effects than any action on intestinal motility and pro- pulsion (Oben et al., 2006). Also, tannins partially cover the mucus membrane of the gut and thereby a barrier against toxins Table 4 is established with an inhibiting effect on diarrhea in humans Antibacterial activity of EtOH (70%) extract, its fractionations and pure (Chung et al., 1998). compounds isolated from the EtOAc extract of Hydnora johannis roots The antibacterial activity of water extract devoid of tannins expressed as MICs (μg/mL). and free tannin fraction was evaluated and results of the MIC Extracts E. E. coli S. S. P. values were presented in Table 3. Although the aqueous extract faecalis aureus aureus 2 aeruginosa showed significant activity against E. faecalis (MIC value of ETOH (70%) 16 N256 64 64 N256 16 μg/mL) and S. aureus strains (MIC value of 64 μg/mL), Hexane N256 N256 N256 N256 N256 N N N N N the water extract without tannins fraction as well as the separat- CHCl3 256 256 256 256 256 N N N ed tannin fraction did not show any activity against all the test- EtOAc 256 256 128 128 256 WAFHCE N256 N256 128 128 N256 ed organisms. The decrease in potency of the water extracts Catechin ND ND N256 N256 ND when subjected to different fractions seemed to indicate the Vanillin ND ND N256 N256 ND loss of synergistic action between any of the phytochemical Protocatechuic acid ND ND N256 N256 ND constituents present in the water due to its subjection to the sep- WAFHCE, water after fractionation with hexane, CHCl3 and EtOAc; MICs aration process. When a number of compounds are acting values are representative of three independent determinations. S. Yagi et al. / South African Journal of Botany 78 (2012) 228–234 233

Fig. 1. Structural formulae of the isolated compounds from Hydnora johannis roots. (2), trans 3′5-dihydroxy-4′7-dimethoxydihydroflavonol; (4), vanillin; (5), protocatechuic acid; (6), catechin. diffusion method was carried out by Elegami et al. (2001) and might exert moderate cytotoxic effect against KB cells but not Saadabi and Ayoub (2009). In their results all extracts inhibited on normal MRC5 cells, while none of the tested phenolic com- the growth of tested bacteria. However, this difference might ponents isolated from the 70% EtOH extract demonstrated any be in relation to the high concentration of extracts used sign of toxicity at the highest concentration tested and no syn- (100 mg/mL) by the former authors. Saadabi and Ayoub (2009) ergism was observed on different combination of pure com- did not give details of the concentrations used which make the pounds. Thus the moderate cytotoxicity of extracts may have comparison of their results to ours difficult. been contributed by other components in the mixture.

3.3. Chemical constituents 4. Conclusions

The EtOAc fraction was further investigated for its phyto- The results obtained strongly suggest that the phenolics are chemical characteristics. Six compounds were isolated and their the major components of this plant and H. johannis could be structures were determined by NMR and mass fragmentation a potential source of natural bioactive chemicals like catechins, and then confirmed by comparison with the literature data and protocatechuic acid and vanillin. with authentic specimens as cirsiliol (3′,4′,5-trihydroxy-6-7- This work also permitted the evaluation of antibacterial activity dimethoxy-flavone) (1)(Abdelshafeek et al., 2009), trans 3′5- and cytotoxicity of H. johannis roots extracts. The water dihydroxy-4′7-dimethoxydihydroflavonol (2)(Islam and Tahara, extract could be considered as potent antimicrobial agents against 2000), oleic acid (3), vanillin (4-hydroxy-3-methoxybenzalde- E. faecalis, B. subtilis and S. aureus. Although the water extract hyde) (4)(Yu et al., 2006), protocatechuic acid (3,4 dihydroxy is used in the treatment of diarrhea and dysentery in the folk med- benzoic acid) (5)(Yu et al., 2006) and dl catechin (trans (+−)2- icine, the extract did not possess any activity against bacteria (3,4-dihydroxyphenyl-3,4-dihydro-2H-1-benzopyran-3,5,7- spp. that are mainly responsible for diarrhea (Escherichia spp., Sal- triol)) (6)(Du et al., 2005). Four compounds were also identified monella spp., Shigella spp., or Yersinia spp). This leads to the by GCMS as stigmasterol (7), oleic acid (3), myristic acid (8), palmitic acid (9)(Fig. 1). Copies of the original spectra are ob- Table 5 tainable from the corresponding author. Cytotoxic effect of extracts and pure compounds from Hydnora johannis roots against KB cells lines. 3.4. Determination of cytotoxicity Extract/ Concentrations KB cell lines compound Inhibition (%) IC50 (μg/mL) Cytotoxic effect of the water and EtOH (70%) extracts Water 10 μg/mL 65±3 N50 against MRC5 and KB cells lines was carried out and the cyto- EtOH (70%) 10 μg/mL 41±5 N50 toxicities of compounds 1–4 and mixture were also evaluated (4)10− 5 M and 10− 6 MNt – against KB cell lines (Table 5). At concentration of 10 μg/mL (5)10− 5 M and 10− 6 MNt – − − the water and EtOH extracts showed moderate toxicity with (6)105 M and 10 6 MNt – (4)+(5)10− 5 M/10−5 MNt – 41%±5 and 65%±3 respectively while they were devoid of − − (4)+(6)105 M/10 5 MNt – μ − − any toxicity at concentration of 1 g/mL. Both extracts at 10 (5)+(6)105 M/10 5 MNt – and 1 μg/mL did not show any cytotoxic activity against (4)+(5)+(6)10− 5 M/10−5 M/10− 5 MNt – MRC5 cells. The pure compounds and their mixture did not Taxotere 93.0±0.3 – exert an inhibitory effect at the concentrations tested. The (4), catechin; (5), vanillin, (6), protocatechuic acid; Nt, not toxic; values are rep- data reported here thus suggest that water and EtOH extracts resentative of two independent determinations. 234 S. Yagi et al. / South African Journal of Botany 78 (2012) 228–234 suggestion that, the efficacy of this plant in the treatment of this dis- Mahdi, H., Palmina, P., Glavtch, I., 2006. Characterization of Acacia nilotica an ease was based on the physical and chemical properties of tannins. indigenous tanning material of Sudan. Journal of Tropical Forest Science 18, 181–187. Little cytotoxicity is reported against the KB cell lines. The safety Meyrowitsch, D.W., Bygbjerg, I.b.C., 2007. Global burden of disease — a race of this plant in the traditional medicine should be verified by much against time. Danish Medical Bulletin 54, 32–34. further testing, including in vivo experiments and clinical studies. Miranda, D., Pereira, L., Sirsat, S.M., Antarkar, D.S., Vaidya, A.B., 1993. In vitro action of Dadima (Punica granatum Linn.) against microorganisms in- — Acknowledgments volved in human gastrointestinal infections isolation and identification of tannins. Journal of Research in Ayurveda and Siddha 14, 154–164. Murray, C.J.L., Lopez, A.D., Mathers, C.D., Stein, C., 2001. The Global Bur- Ms S. Yagi is thankful to Prof. B.M. Abegaz (Department of den of Disease 2000 Project: Aims, Methods, and Data Sources. World Chemistry, Faculty of Science, University of Botswana, Gaberone, Heath Organization, Geneva. Botswana) for spectral analysis of some of the isolated compounds. Musselman, L.J., 1984. Parasitic angiosperms of Sudan: Orobanchaceae, Hyd- noraceae, and Cuscuta. Notes Royal Botanic Garden, Edinburgh 42, 21–39. Musselman, L.J., 1991. The genus Hydnora (Hydnoraceae). Proceedings of the References 5th International Symposium of Parasitic Weeds. CIMMYT, Nairobi, Kenya, pp. 247–250. Abdelshafeek, K.A., Abdelrahem, F., Elwahsh, M.A., Abdelkhalek, I.A., 2009. In- Musselman, L.J., Visser, J.H., 1987. Hydnora johannis in Southern Africa. Din- vestigation of the flavonoidal constituents and insecticidal activity of Teucrium teria 19, 77–82. zanonii. Pharmacognosy Research 1, 410–416. Musselman, L.J., Visser, J.H., 1989. Taxonomy and natural history of Hydnora Agunu, A., Yusuf, S., Andrew, G.O., Zezi, A.U., Abdurahman, E.M., 2005. (Hydnoraceae). Aliso 12, 317–326. Evaluation of five medicinal plants used in diarrhoea treatment in Nigeria. National Committee for Clinical Laboratory Standards, 2003. Methods for Dilu- Journal of Ethnopharmacology 101, 27–30. tion Antimicrobial Susceptibility Tests for Bacteria that Grow Aerobically; Akiyama, H., Fujii, K., Yamasaki, O., Oono, T., Iwatsuki, K., 2001. Antibacter- Approved Standard M7-A6. NCCLS, Wayne, Pa, USA. ial action of several tannins against Staphylococcus aureus. Journal of An- Nikaido, H., 1994. Prevention to drug access to bacterial targets: permeability – timicrobial Chemotherapy 48, 487 491. barriers and active efflux. Science 264, 382–388. Ambasta, S.P., 1994. The useful plants of India. Publication and Information Oben, J.E., Assi, S.E., Agbor, G.A., Musoro, D.F., 2006. Effect of Eremomastax Directorate. CSIR, New Delhi, p. 4. speciosa on experimental diarrhoea. African Journal. Traditional. Comple- Atta, A.H., Mouneir, S.M., 2004. Antidiarrhoeal activity of some Egyptian me- mentary and Alternative Medicines 3, 95–100. dicinal plant extracts. Journal of Ethnopharmacology 92, 303–309. Ordon Ez, A.A.L., Gomez, J.D., Vattuone, M.A., Isla, M.I., 2006. Antioxidant ac- Ayoub, S.M.H., 1984. Flavanol molluscides from the Sudan Acacias. Crude tivities of Sechium edule (Jacq.) Swart extracts. Food Chemistry 97, 452–458. Drug Research 232, 87–90. Ryabushkina, N.A., 2005. Synergism of metabolite action in plant responses to Bolin, J.F., Maass, E., Musselman, L.J., Tennakoon, K.U., 2005. Floral biology of stresses. Russian Journal of Plant Physiology 52, 547–552. three Namibian Hydnora (Hydnoraceae). Proceedings of the Botanical Society Saad, B., Azaizeh, H., Abu-Hijleh, G., Said, O., 2006. Safety of traditional Arab of America, Austin, Texas, pp. 56–57http://www.odu.edu/~lmusselm/plant/ herbal medicine. Evidence-based Complementary and Alternative Medicine parasitic/hydnora_liter.php. 3, 433–439. Brunton, L.L., 1996. Agents for control of gastric acidity and treatment of pep- Saadabi, A.M.A., Ayoub, S.M.H., 2009. Comparative bioactivity of Hydnora tic ulcers, Goodman and Gilman's the Pharmacological Basis of Therapeu- abyssinica A. Braun against different groups of fungi and bacteria. Journal – tics, 9th ed. McGraw-Hill, New York, pp. 901 915. of Medicinal Plants Research 3, 262–265. Cannas, A., 1999. Tannins. Toxic and anti-nutritional effects. Phytochemistry Scherer, R., Gerhardt, P., 1971. Molecular sieving by the Bacillus megaterium 22, 1049. cell wall and protoplast. Journal of Bacteriology 107, 718–735. Casburn-Jones, A.C., Farthing, M.J., 2004. Management of infectious diar- Snyder, J.D., Merson, M.H., 1982. The magnitude of the global problem of rhoea. Gut 53, 296–305. acute diarrhoeal disease: a review of active surveillance data. Bulletin of CA-SFM, 2008. Communiqué 2008 (Edition janvier 2008). http://www.sfm. the World Health Organisation 60, 604–613. asso.fr/publi/general.php?pa=12008. Sun, J.S., Tsuang, Y.W., Chen, I.J., Huang, W.C., Hang, Y.S., Lu, F.J., 1998. Chung, K., Wong, T.Y.Y., Wei, C., Huang, Y., Lin, Y., 1998. Tannins and An ultra-weak chemiluminescence study on oxidative stress in rabbits fol- human health: a review. Critical Reviews in Food Science and Nutrition lowing acute thermal injury. Burns 24, 225–231. 38, 421–464. Tennakoon, K.U., Bolin, J.F., Musselman, L.J., Maass, E., 2007. Structural at- Devi, B.P., Boominathan, R., Mandal, S.C., 2002. Evaluation of anti-diarrhoeal tributes of the hypogeous holoparasite Hydnora triceps Drège & Meyer – activity of Cleome viscosa L. extract in rats. Phytomedicine 9, 739 742. (Hydnoraceae). American Journal of Botany 94, 1439–1449. Dold, T., Cocks, M., 2003. Fine fare, rare remedy. Veld and Flora 89, 12–14. Williams, V.L., Wojtasik, E.M., Witkowski, E.T.F., 2011a. Ethno-ecological Du, Z., Zhao, P., He, H., Zhu, N., Hao, X., Shen, Y., 2005. The chemical con- evidence for Hydnora abyssinica occurring in Johannesburg and Durban stituents from the callus culture of Trewia nudiflora. Helvetica Chimica traditional medicine markets. South African Journal of Botany 77, 268–279. Acta 88, 2424–2429. Williams, V.L., Falcão, M.P., Wojtasik, E.M., 2011b. Hydnora abyssinica: eth- El Ghazali, G.E., 1997. The Promising Medicinal Plants of the Sudan. National nobotanical evidence for its occurrence in southern Mozambique. South Af- Council for Research. Khartoum University Press, Sudan. rican Journal of Botany 77, 474–478. Elegami, A.A., Almagboul, A.Z., Omer, A.E.A., El Tohami, M.S., 2001. Suda- Wolfe, K., Wu, X., Liu, R.H., 2003. Antioxidant activity of apple peels. Journal nese plants used in folkloric medicine: screening for antibacterial activity. of Agricultural and Food Chemistry 51, 609–614. Part X. Fitoterapia 72, 810–817. Xiao, J.B., Jiang, X.Y., Chen, X.Q., 2005. Antibacterial, anti-inflammatory and Fontanay, S., Grare, M., Mayer, J., Finance, C., Duval, R.E., 2008. Ursolic, diuretic effect of flavonoids from Marchantia convolute. African Journal of oleanolic and betulinic acids: antibacterial spectra and selectivity indexes. Traditional, Complementary, and Alternative Medicines 2, 244–252. Journal of Ethnopharmacology 120, 272–276. Yu, L.L., Liao, J.F., Chen, C.F., 2000. Anti-diarrhoeal effect of water extract of Hutchings, A., Scott, A.H., Lewis, G., Cunningham, A.B., 1996. Zulu Medici- Evodiae fructus in mice. Journal of Ethnopharmacology 73, 39–45. nal Plants. An Inventory. University of Natal Press, Pietermaritzburg. Yu, Y., Gao, H., Tang, Z., Song, X., Wu, L., 2006. Several phenolic acids from the Islam, Md.T., Tahara, S., 2000. Dihydroflavonols from Lannea coromandelica. fruit of Capparis spinosa. Asian Journal of Traditional Medicines 1, 1–4. Phytochemistry 54, 901–907.

Edited by SV Vuuren